Shell element (deforms good but no force transmitted)

[Anonymous]

Hi, 

I am trying to simulate a simple 3 ply laminate composite plate (500x275) on a 3-point bend test setup.
I tried the same test with a steel plate in place, just to see if everything works and it does beautifully, all the stresses, strains and reaction vectors look good. 
Then I made a midplane surface from that plate to use in shell element analysis. I kept all the constraints equal as before, just modifying the contact surfaces to compensate for the change. (I also used separated contacts for support structure, bonded for penetrator, and specified required max activation distance (slightly bigger than the gap).
I specified all the plies and materials that represent them as 2D orthotropic, gave them the required properties, so I don't think I made an error there.

The materials for penetrator and supports are just steel with a bit higher Young modulus than usual.

The problem is that the shell structure (test specimen) shows good deformation in results, but when I choose any type of stress contour, it shows nothing on the test specimen and support structures. Only the composites data shows some green areas on the test specimen.

I also tried using offset bonded contacts and I do get some reaction forces and stresses in the support structure, but that kind of support is not suitable for my analysis.
It looks like it doesn't want to transmit forces in any other type of bond.

Do I need any more constraints maybe? I fixed the supports, penetrator can only move in Y direction, and I applied Z symmetry on faces of penetrator and support structures. (my coordinate system is a bit off, yes)

Here are some screenshots

setupno stressesdisplacementscomposite stressesoffset bonded

 Thanks

[John_Holtz]

Hi @Anonymous

I have looked at your images, but I am not entirely clear what problem you are seeing. In the image that says "no stress", there is stress. Are you saying that the stress on the supports and penetrator is smaller than you expected? If that is the problem, how much force is required to move the penetrator? If it is smaller than you expect, that would help to determine the problem. (You can right-click on one of the constraints and then choose "SPC Summation". From that dialog, you can choose which constraint to view the support reactions.)

I think there is a good chance that your model is not statically stable. The shell is probably not in contact with the supports, so the solver is automatically constraining the shell to prevent it from moving like a rigid body. With the artificial constraints, It is then able to solve the analysis, but then the contact with the supports is not correct. (Because of the midplane idealization, there is a gap between the shell and the supports, so the parts are not in contact until a load is applied, but the solver cannot calculate the displacement if they are not in contact.)

I suggest that you try using a nonlinear static analysis. It can do a better job of "closing the gap" between the parts. But even then, you may need to add springs to provide some static stability to the shell in the vertical direction.

Let us know what you find out.

[Anonymous]

Thanks for your reply,

I agree it is most probably a bad constraint somewhere.
Could you run through these screenshots of the details and tell me if I'm missing something?

Besides the costraints below I also constrained (locked) the lower surface of both support pins in all travels and rotation.

The contacts are: 2 separation contacts with max activation distance higher than the actual distance and 1 bonded contact (loading pin to surface).

Thanks

enforced motion (y is the vertical axis)loading pin (only in vertical)shell surface edges restricted in x and zresults